[Seminar] From Neural Principles to Machine Intelligence by Prof. Se-Bum Paik, KAIST

Prof. Se-Bum Paik, Department of Brain and Cognitive Sciences, KAIST

Title: ​ From Neural Principles to Machine Intelligence

Abstract:
How do cognitive functions emerge from neural circuits, and what can the brain teach us about building intelligent machines? Despite remarkable advances in artificial intelligence, current models remain heavily dependent on large-scale training data and often struggle to generalize beyond their training distributions, limiting their robustness to diverse and changing environmental conditions. Understanding how biological neural circuits give rise to cognitive functions may offer new insights into developing more data-efficient, adaptive, and robust artificial intelligence. This talk will explore how fundamental principles of neural development and organization can inspire alternative approaches to artificial intelligence. From understanding how intrinsically organized, hard-wired circuits in the early visual system support environment-independent perception and robust continual learning to developing hierarchical learning architectures inspired by human cognitive development, these findings demonstrate the potential of biological principles to reshape artificial intelligence. Moving beyond the conventional paradigm of learning everything from data may open new avenues for building intelligent machines that are not only efficient and robust, but also more closely reflect the principles of the human brain.